SmartPipeline 0.7.1

Install

Install from PyPI, no dependencies will be installed:

pip install smartpipeline

Writing your pipeline

SmartPipeline is designed to help the developer following best practices, the design is based on industrial experience on data products.

SmartPipeline focuses on simplicity and efficiency in handling data locally, i.e. serialization and copies of the data are minimized.

Main features:

• Define a pipeline object as a sequence of stateful stage objects, optionally set a source on which the pipeline iterates.

• A pipeline can run indefinitely on the source or it can be used to process single items.

• Stages can run concurrently on the source or on single asynchronously processed items.

• A stage can be designed for processing batches, i.e. sequences of consecutive items, at once.

• Custom error handling can be set for logging and monitoring at stage level.

An example of a trivial pipeline definition and run:

class RandomGenerator(Source):
    def pop(self):
        item = Item()
        item.data["number"] = random.random()
        return item


class Adder(Stage):
    def __init__(self, value):
        self.value = value

    def process(self, item):
        item.data["number"] += self.value
        return item


class Rounder(BatchStage):
    def process_batch(self, items):
        for item in items:
            item.data["number"] = round(item.data["number"], 1)
        return items


pipeline = (
    Pipeline()
    .set_source(RandomGenerator())
    .append("adder", Adder(1), concurrency=2)
    .append("rounder", Rounder(size=100))  # process 100 items at a time
    .build()
)


for item in pipeline.run():
    print(item)

Read the documentation for an exhaustive guide.

The examples folder contains full working sample pipelines.

Future improvements:

• Stages can be memory profiled.

• Processed items can be cached at stage level.